S
Sophia Kirova
Guest
Progress MarkLogic now supports connecting AI assistants like Claud, AWS Quick and Copilot Studio to your data to empower your users with the answers they need in seconds.
Built on the innovations introduced in MarkLogic Server 12, this release helps you operationalize enterprise knowledge with AI agents and introduces new capabilities to scale growing data and AI workloads more flexibly and cost-effectively, locally and in the cloud, through new capabilities that optimize storage, improve cloud elasticity and reduce infrastructure costs.
Key enhancements include:
AI agents have been the defining 2026 moment and an opportunity speeding up innovation cycles and time to value from complex, time-intense tasks. AI agents also remove significant operational overhead and shorten timelines when it comes to bringing information to decision-makers and subject matter experts.
The MarkLogic MCP Server allows AI agents to answer user questions on your trusted data in natural language, accelerating agentic AI initiatives. It gives AI agents secure, governed access to enterprise knowledge stored in MarkLogic Server through a standardized and extensible retrieval layer.
You can connect AI assistants like Copilot, AWS Quick and Claude directly to MarkLogic Server so they can securely discover, retrieve, reason over and enrich enterprise content while preserving MarkLogic's security, provenance and semantic context.
You can also add MarkLogic-stored documents and collections to your orchestration pipelines and agentic multi-step workflows, so relevant content can be leveraged by agents in the process.
For example, if you wanted to use policy guidelines stored in MarkLogic Server to be used as a checkpoint in report generation to automatically check for compliance gaps, you can create a custom tool in the MCP Server that will be discovered and used by an agent.
Out of the box, the MarkLogic MCP Server is database-context aware and understands database schema and configuration. Extensible by design, the MCP Server allows you to build your own custom tools on top of MarkLogic MCP Server functionality, define resources and custom prompts.
To get started with the MarkLogic MCP Server, download it from our GitHub repository and watch this quick start guide to get all set up.
As your AI applications scale, so do your operational costs. We wanted to give you more control over your retrieval strategy and are delivering new flexibility enhancements for you to keep growing search overhead under control.
The new variable-precision vector compression allows you to optimize vector storage and indexing costs by choosing the level of retrieval precision your workloads require. You can control vector precision at the schema level without changing application code to reduce the footprint of large-scale vector and hybrid searches.
MarkLogic Server 12.1 also adds flexible indexing controls that allow developers to exclude parts of documents from the universal index. That way you can maximize storage efficiency by indexing only the content that delivers search value. If you manage large amounts of content, such as vector embeddings, generated metadata or encoded files, you don’t need to incur the infrastructure and operational costs of indexing them in their entirety.
Needless to say, those enhancements will also significantly improve hybrid query performance. To boost performance even further, we've extended support for parameterized CTS queries in Optic API so you can build complex hybrid queries with more flexible, reusable query logic. You can change search terms, filters or structured conditions by passing parameters instead of rebuilding query strings. This lets you cache Optic plans for better performance and avoid error-prone string‑based query construction.
Cloud spend represents a significant portion of your total cost of ownership. If you're running MarkLogic Server on AWS Cloud, you can reduce your cloud bill by switching to AWS Graviton processors.
Support for AWS Graviton instances in MarkLogic Sercer 12.1 gives you a more cost-effective way to run MarkLogic workloads on AWS Cloud. Built on the Arm-based architecture, Graviton processors are designed to deliver better price-performance than comparable x86-based infrastructure. You can support the same or greater throughput at a lower compute cost, helping you free up budget for more memory or storage while maintaining the enterprise-grade reliability you expect from MarkLogic.
By leveraging AWS’s architecture efficiencies, you can future ready your enterprise foundation to run data-intensive and AI-driven applications more economically on the modern cloud infrastructure AWS is prioritizing for the next decade.
We first introduced support for AWS Graviton with the 11.3 LTS release, and our team has been continually working on significant performance improvements in MarkLogic Server 12.1. Our tests show that Graviton's multi-threaded architecture drives gains in parallelizable workloads, including geospatial, API queries and load operations. Compared to existing x86 architectures, you can expect higher performance under concurrent multi-threaded load, especially if you have more than 90 users or clients trying to access data in MarkLogic simultaneously.
Interested in evaluating your price-performance gains? You can book a consultation with our experts to test your workloads. Our team can provide a clear and reliable path forward.
We've now made it easy to expand storage, scale capacity on demand and maintain high availability in the cloud.
To help you support consistent performance and business continuity during periods of high application usage, the MarkLogic Server 12 release introduced dynamic hosts that allow administrators to scale clusters on demand. You can now declaratively add clusters to handle peak query workloads through the MarkLogic Kubernetes Operator and AWS CloudFormation Template.
The MarkLogic Kubernetes Operator also supports increasing storage capacity for MarkLogic clusters by updating the custom resource. It validates resize prerequisites, expands PVCs for primary and additional persistent volumes, handles both online and restart-required filesystem expansion paths and reports progress through status and events.
And when you are ready, you can upgrade to the latest version of MarkLogc Server with zero downtime. Rolling upgrades in the MarkLogic Kubernetes Operator allow you to move to supported MarkLogic releases without disrupting production environments.
Join an upgrade workshop and watch the release webinar for live demos, additional updates we have not covered here and the opportunity to ask questions.
Continue reading...
Built on the innovations introduced in MarkLogic Server 12, this release helps you operationalize enterprise knowledge with AI agents and introduces new capabilities to scale growing data and AI workloads more flexibly and cost-effectively, locally and in the cloud, through new capabilities that optimize storage, improve cloud elasticity and reduce infrastructure costs.
Key enhancements include:
- The new MarkLogic MCP Server complements the intelligent AI search capabilities from MarkLogic Server 12 to enable governed agentic retrieval
- Search and indexing enhancements allow you to optimize operational cost by balancing search accuracy with memory and storage efficiency
- The Virtual Views in-line enhancement gives you the control and flexibility to instantly generate relational views only for the duration of the query without extra admin configuration
- Support for AWS Graviton coupled with powerful scaling enhancements complement modern cloud readiness and elasticity, making MarkLogic applications easier and more cost-effective to run in cloud environments
Activate Agentic AI on Your MarkLogic Data
AI agents have been the defining 2026 moment and an opportunity speeding up innovation cycles and time to value from complex, time-intense tasks. AI agents also remove significant operational overhead and shorten timelines when it comes to bringing information to decision-makers and subject matter experts.
The MarkLogic MCP Server allows AI agents to answer user questions on your trusted data in natural language, accelerating agentic AI initiatives. It gives AI agents secure, governed access to enterprise knowledge stored in MarkLogic Server through a standardized and extensible retrieval layer.
You can connect AI assistants like Copilot, AWS Quick and Claude directly to MarkLogic Server so they can securely discover, retrieve, reason over and enrich enterprise content while preserving MarkLogic's security, provenance and semantic context.
You can also add MarkLogic-stored documents and collections to your orchestration pipelines and agentic multi-step workflows, so relevant content can be leveraged by agents in the process.
For example, if you wanted to use policy guidelines stored in MarkLogic Server to be used as a checkpoint in report generation to automatically check for compliance gaps, you can create a custom tool in the MCP Server that will be discovered and used by an agent.
Out of the box, the MarkLogic MCP Server is database-context aware and understands database schema and configuration. Extensible by design, the MCP Server allows you to build your own custom tools on top of MarkLogic MCP Server functionality, define resources and custom prompts.
To get started with the MarkLogic MCP Server, download it from our GitHub repository and watch this quick start guide to get all set up.
Scale AI and Hybrid Search More Efficiently
As your AI applications scale, so do your operational costs. We wanted to give you more control over your retrieval strategy and are delivering new flexibility enhancements for you to keep growing search overhead under control.
The new variable-precision vector compression allows you to optimize vector storage and indexing costs by choosing the level of retrieval precision your workloads require. You can control vector precision at the schema level without changing application code to reduce the footprint of large-scale vector and hybrid searches.
MarkLogic Server 12.1 also adds flexible indexing controls that allow developers to exclude parts of documents from the universal index. That way you can maximize storage efficiency by indexing only the content that delivers search value. If you manage large amounts of content, such as vector embeddings, generated metadata or encoded files, you don’t need to incur the infrastructure and operational costs of indexing them in their entirety.
Needless to say, those enhancements will also significantly improve hybrid query performance. To boost performance even further, we've extended support for parameterized CTS queries in Optic API so you can build complex hybrid queries with more flexible, reusable query logic. You can change search terms, filters or structured conditions by passing parameters instead of rebuilding query strings. This lets you cache Optic plans for better performance and avoid error-prone string‑based query construction.
Get Better Price-Performance for AWS Cloud
Cloud spend represents a significant portion of your total cost of ownership. If you're running MarkLogic Server on AWS Cloud, you can reduce your cloud bill by switching to AWS Graviton processors.
Support for AWS Graviton instances in MarkLogic Sercer 12.1 gives you a more cost-effective way to run MarkLogic workloads on AWS Cloud. Built on the Arm-based architecture, Graviton processors are designed to deliver better price-performance than comparable x86-based infrastructure. You can support the same or greater throughput at a lower compute cost, helping you free up budget for more memory or storage while maintaining the enterprise-grade reliability you expect from MarkLogic.
By leveraging AWS’s architecture efficiencies, you can future ready your enterprise foundation to run data-intensive and AI-driven applications more economically on the modern cloud infrastructure AWS is prioritizing for the next decade.
We first introduced support for AWS Graviton with the 11.3 LTS release, and our team has been continually working on significant performance improvements in MarkLogic Server 12.1. Our tests show that Graviton's multi-threaded architecture drives gains in parallelizable workloads, including geospatial, API queries and load operations. Compared to existing x86 architectures, you can expect higher performance under concurrent multi-threaded load, especially if you have more than 90 users or clients trying to access data in MarkLogic simultaneously.
Interested in evaluating your price-performance gains? You can book a consultation with our experts to test your workloads. Our team can provide a clear and reliable path forward.
Enjoy Better Cloud Elasticity
We've now made it easy to expand storage, scale capacity on demand and maintain high availability in the cloud.
To help you support consistent performance and business continuity during periods of high application usage, the MarkLogic Server 12 release introduced dynamic hosts that allow administrators to scale clusters on demand. You can now declaratively add clusters to handle peak query workloads through the MarkLogic Kubernetes Operator and AWS CloudFormation Template.
The MarkLogic Kubernetes Operator also supports increasing storage capacity for MarkLogic clusters by updating the custom resource. It validates resize prerequisites, expands PVCs for primary and additional persistent volumes, handles both online and restart-required filesystem expansion paths and reports progress through status and events.
And when you are ready, you can upgrade to the latest version of MarkLogc Server with zero downtime. Rolling upgrades in the MarkLogic Kubernetes Operator allow you to move to supported MarkLogic releases without disrupting production environments.
Join an upgrade workshop and watch the release webinar for live demos, additional updates we have not covered here and the opportunity to ask questions.
Continue reading...